Reserch On Urban Traffic Coordinated Control Based On Multi-Agent

With the development of social economy, the urban road traffic pressure and congestion have been under a severe strain, these problems caused universal social attention. However people increase the city road coverage by building more roads,it cannot solve the traffic demand. We must take the high technology into the transport system, so as to achieve vastly improved network communication ability and service quality goal. Based on the study and analysis of the classic and today's some traffic signal control system, analyzes the premise of traffic control of the development direction, found the traditional depend on mathematical model control method, and the limitations of urban traffic flow doesn't fit for our motor vehicles and pedestrians more than the actual. For solving China's modern urban traffic problem,being outstanding constantly introducing artificial intelligence technology, intellectualizdthe traffic signal control system.It firstly introduces urban traffic signal control system's history and status in the domestic and overseas, the traffic and basic theoretical knowledge. For building a city in intelligent transport system, this paper studies the important basic theory method does support, including Agent theory, knowledge base theory, production system theory and Inference Engine strategy. On this basis, the introduction of the Agent concept, put forward and recounted the frame and controlling strategy of Signal Coordinated Control Based On Multi-Agent System. Bold attempt will produce type system into Agent model, with a new type of high intelligented body as part of a system for various algorithms, it provides body and action for decision-making, empirical knowledge. Meanwhile this paper designed several important business algorithms and model of evaluation indexes, including signal timing model and algorithm, the headway distribution pattern matching algorithm with simulation atMatlab, traffic flow sequence segmentation algorithm, traffic flow predictive model and simulation at Matlab, etc.Through this paper, the research provides some new ideas and methods for our country's current intelligenttransport system research; it provided many reference to signal timing model algorithm; it applied certain reference valueto multi-agent, knowledge base, production rules and Inference Engine theory.